AWAKE, the Advanced Proton Driven Plasma Wakefield Acceleration Experiment at CERN

Home , AWAKE, CERN, CTF3 (CERN), Proton Synchrotron

Nuclear Instruments and Methods in Physics Research A ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect Nuclear Instruments and Methods in Physics Research A

journal homepage: www.elsevier.com/locate/nima

AWAKE, The Advanced Proton Driven Plasma Wakeﬁeld Acceleration Experiment at CERN

E. Gschwendtner b, E. Adli v, L. Amorim g, R. Apsimon c,j, R. Assmann e, A.-M. Bachmann l, F. Batsch l, J. Bauche b, V.K. Berglyd Olsen v, M. Bernardini b, R. Bingham o, B. Biskup b,d, T. Bohl b, C. Bracco b, P.N. Burrows i,w, G. Burt c, B. Buttenschön m, A. Butterworth b, A. Caldwell l, M. Cascella s, E. Chevallay b, S. Cipiccia x, H. Damerau b, L. Deacon s, P. Dirksen r, S. Doebert b, U. Dorda e, J. Farmer f, V. Fedosseev b, E. Feldbaumer b, R. Fiorito c,t, R. Fonseca g, F. Friebel b, A.A. Gorn a,n, O. Grulke m, J. Hansen b, C. Hessler b,W.Hoﬂe b, J. Holloway i,w, M. Hüther l,q, D. Jaroszynski x, L. Jensen b, S. Jolly s, A. Joulaei l, M. Kasim i,w, F. Keeble s, Y. Li c,u, S. Liu r, N. Lopes h,g, K.V. Lotov a,n, S. Mandry s, R. Martorelli f, M. Martyanov l, S. Mazzoni b, O. Mete c,u, V.A. Minakov a,n, J. Mitchell c,j, J. Moody l, P. Muggli l, Z. Najmudin h,i, P. Norreys w,o,E.Özl, A. Pardons b, K. Pepitone b, A. Petrenko b, G. Plyushchev b,p, A. Pukhov f, K. Rieger l,q, H. Ruhl k, F. Salveter b, N. Savard l,r,y, J. Schmidt b, A. Seryi i,w, E. Shaposhnikova b, Z.M. Sheng x, P. Sherwood s, L. Silva g, L. Soby b, A.P. Sosedkin a,n, R.I. Spitsyn a,n, R. Trines o, P.V. Tuev a,n, M. Turner b, V. Verzilov r, J. Vieira g, H. Vincke b,Y.Weic,t, C.P. Welsch c,t, M. Wing s,e, G. Xia c,u, H. Zhang c,t a Budker Institute of Nuclear Physics SB RAS, Novosibirsk 630090, Russia b CERN, Geneva, Switzerland c Cockcroft Institute, Warrington WA4 4AD, UK d Czech Technical University, Zikova 1903/4, 166 36 Praha 6, Czech Republic e DESY, Notkestrasse 85, 22607 Hamburg, Germany f Heinrich-Heine-University of Düsseldorf, Moorenstrasse 5, Düsseldorf 40225, Germany g GoLP/Instituto de Plasmas e Fusão Nuclear, Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal h John Adams Institute for Accelerator Science, Blackett Laboratory, Imperial College London, London SW7 2BW, UK i John Adams Institute for Accelerator Science, Oxford, UK j Lancaster University, Lancaster LA1 4YR, UK k Ludwig-Maximilians-Universität, Munich 80539, Germany l Max Planck Institute for Physics, Föhringer Ring 6, München 80805, Germany m Max Planck Institute for Plasma Physics, Wendelsteinstr. 1, Greifswald 17491, Germany n Novosibirsk State University, Novosibirsk 630090, Russia o STFC Rutherford Appleton Laboratory, Didcot OX11 0QX, UK p Swiss Plasma Center, EPFL, Lausanne 1015, Switzerland q Technische Universität München (TUM), Arcisstrasse 21, D-80333 Munich, Germany r TRIUMF, 4004 Wesbrook Mall, Vancouver V6T2A3, Canada s UCL, Gower Street, London WC1E 6BT, UK t University of Liverpool, Liverpool L69 7ZE, UK u University of Manchester, Manchester M13 9PL, UK v University of Oslo, Oslo 0316, Norway w University of Oxford, Oxford OX1 2JD, UK x University of Strathclyde, 16 Richmond Street, Glasgow G1 1XQ, UK y University of Victoria, 3800 Finnerty Rd, Victoria, Canada

http://dx.doi.org/10.1016/j.nima.2016.02.026 0168-9002/& 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Please cite this article as: E. Gschwendtner, et al., Nuclear Instruments & Methods in Physics Research A (2016), http://dx.doi.org/ 10.1016/j.nima.2016.02.026i 2 E. Gschwendtner et al. / Nuclear Instruments and Methods in Physics Research A ∎ (∎∎∎∎) ∎∎∎–∎∎∎ article info abstract

Article history: The Advanced Proton Driven Plasma Wakefield Acceleration Experiment (AWAKE) aims at studying Received 30 November 2015 plasma wakefield generation and electron acceleration driven by proton bunches. It is a proof-of- Received in revised form principle R&D experiment at CERN and the world's first proton driven plasma wakefield acceleration 5 February 2016 experiment. The AWAKE experiment will be installed in the former CNGS facility and uses the 400 GeV/c Accepted 9 February 2016 proton beam bunches from the SPS. The first experiments will focus on the self-modulation instability of the long (rms 12 cm) proton bunch in the plasma. These experiments are planned for the end of 2016. Keywords: Later, in 2017/2018, low energy (15 MeV) electrons will be externally injected into the sample wake- AWAKE fields and be accelerated beyond 1 GeV. The main goals of the experiment will be summarized. A Proton driven plasma wakefield accelera- summary of the AWAKE design and construction status will be presented. tion & 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license Linear accelerators Plasma wakefield (http://creativecommons.org/licenses/by/4.0/). Electron acceleration

1. Introduction transverse components of the plasma wakeﬁelds and the wake- ﬁelds being driven by regions of different bunch densities. The

AWAKE is a proof-of-concept acceleration experiment with the modulation period sffiλpe and the modulated bunch resonantly aim to inform a design for high energy frontier particle accel- drives the plasma wakefields. The occurrence of the SMI can be erators and is currently being built at CERN [1,2]. The AWAKE detected by characterizing the longitudinal structure of the proton experiment is the world's first proton driven plasma wakefield beam when exiting the plasma cell. acceleration experiment, which will use a high-energy proton In the AWAKE master schedule, the experiment to obtain evi- bunch to drive a plasma wakefield for electron beam acceleration. dence for the SMI corresponds to Phase 1, and is expected to start A 400 GeV/c proton beam will be extracted from the CERN Super by the end of 2016. In Phase 2, AWAKE aims at the first demon- Proton Synchrotron, SPS, and utilized as a drive beam for wake- stration of proton-driven plasma wakefield acceleration of an fields in a 10 m long plasma cell to accelerate electrons with electron witness beam; this programme is planned to start by the amplitudes up to the GV/m level. Fig. 1 shows the AWAKE facility end of 2017. At a later phase it is foreseen to have two plasma cells in the CERN accelerator complex. In order to drive the plasma in order to separate the modulation of the proton bunch from the wakefields efficiently, the length of the drive bunch has to be on acceleration stage. Simulations [4] show that this would optimize λ the order of the plasma wavelength pe, which corresponds to the acceleration of external electrons and reach even higher E 14– 15 1 mm for the plasma density used in AWAKE (10 10 elec- gradients. trons/cm3). The proton beam for AWAKE has a bunch length of σ ¼ 12 cm, therefore the experiment relies on the self-modulation z 1.1. Baseline design instability (SMI) [3], which modulates the proton driver at the plasma wavelength in the first few meters of plasma. The SMI is a In the baseline design of AWAKE at CERN, an LHC-type proton transverse instability that arises from the interplay between bunch of 400 GeV/c (with an intensity of 3 Â 1011 protons/ bunch) will be extracted from the CERN SPS and sent along the 750 m long proton beam line towards a plasma cell. The AWAKE facility is installed in the area, which was previously used for the CERN Neutrinos to Gran Sasso facility (CNGS) [5]. The proton beam

will be focused to σx;y ¼ 200 μm near the entrance of the 10 m long rubidium vapor plasma cell with an adjustable density in the 1014– 1015 electrons/cm3 range. When the proton bunch, with an r.m.s.

bunch length of σz ¼ 12 cm (0.4 ns), enters the plasma cell, it undergoes the SMI. The effective length and period of the modulated beam is set by the plasma wavelength (for AWAKE, typically

λpe ¼ 1 mm). A high power (E4.5 TW) laser pulse, co-propagating and co-axial with the proton beam, will be used to ionize the neutral gas in the plasma cell and also to generate the seed of the proton bunch self-modulation. An electron beam of 1:2 Â 109 electrons, which will be injected with 10–20 MeV/c into the plasma cell, serves as a witness beam and will be accelerated in the wake of the modulated proton bunch. Several diagnostic tools will be installed downstream of the plasma cell to measure the Fig. 1. CERN accelerator complex. proton bunch self-modulation effects and the accelerated electron

Fig. 2. Baseline design of the AWAKE experiment.

Table 1 combination with a streak camera close to the plasma cell allow AWAKE proton, laser beam and plasma parameters. the overlap and synchronization of the beams to be measured. Optics simulations predict a 1σ spot size of 210 μm at the focal Parameter Baseline point in agreement with the experiment requirements (see Proton beam Table 1). Beam momentum 400 GeV/c Protons/bunch 3 Â 1011 2.2. Electron source Bunch extraction frequency 0.03 Hz (ultimate: 0.14 Hz) Bunch length (σ) 0.4 ns The electron source for AWAKE consists of a 2.5 cell RF-gun and Bunch size at plasma entrance ðσx;yÞ 200 μm Normalized emittance (RMS) 3.5 mm mrad a one meter long booster structure both at 3 GHz (see Fig. 3). The Relative energy spread (Δp/p) 0.035% electron beam is produced via photo-emission by illuminating a ðβn Þ Beta function x;y 4.9 m cathode with a frequency quadrupled laser pulse which is derived ð n Þ 0 Dispersion Dx;y from the main drive laser for the plasma. The wavelength used in Laser beam to plasma the photo injector will be 262 nm. The baseline will use copper fi À 4 Laser type Fibre titanium: sapphire cathodes with a quantum ef ciency of Q e 10 . However, Pulse wavelength (L0) 780 nm thanks to the integration of a load lock system, which allows – Pulse length 100 120 fs transferring cathodes under ultra high vacuum, different cathodes Laser power 4.5 TW À 2 could be used: e.g. Cs2Te with a quantum efficiency of Q 10 . Focused laser size ðσx;yÞ 1mm e Energy stability (RMS) 71.5% A 30 cell travelling wave structure was designed to boost the Repetition rate 10 Hz energy with a constant gradient of 15 MV/m up to a total electron Plasma source energy of 20 MeV. The RF-gun and the booster are powered by a Plasma type Laser ionized rubidium vapor single klystron delivering about 30 MW. The operation mode will Plasma density 7 Â 1014 cm À 3 be single bunch with a maximum repetition rate of 10 Hz. In Length 10 m addition the electron source is equipped with BPMs with a reso- Z Plasma radius 1mm lution of 50 μm to control the beam position, a fast current Skin depth 0.2 mm transformer with a resolution of 10 pC, a Faraday Cup, and two Wavebreaking field E0 ¼ mcωcp=e 2.54 GV/m emittance measurement stations. Details of the electron source are described in [7]. The PHIN Photo-Injector built for CTF3 [8] will be used as the RF-gun for AWAKE. The RF power source will also be recuperated from the CTF3. Table 2 AWAKE electron beam parameters. 2.3. Electron beam line Parameter Baseline Possible range The electron source is installed in an adjacent room 1.16 m Beam energy (MeV) 16 10–20 Energy spread (% ) 0.5 0.5 below the level of the proton beam line. The electron transfer line Bunch length (σ) (ps) 4 0.3–10 consists of an achromatic dog-leg to raise the electron beam up to Beam size at focus (σ) (mm) 250 0.25-1 the level of the proton line and a part which bends the electron Normalized emittance (RMS) (mm mrad) 2 0.5-5 beam horizontally onto the proton beam axis. Details of the beam Charge per bunch (nC) 0.2 0.1À1 line design are described in [9]. Fig. 4 shows the layout of electron beam line, together with the proton and laser beam lines in the bunch properties. Fig. 2 shows the baseline design of the AWAKE AWAKE facility. Directly after the accelerating structure of the experiment. The baseline parameters of the proton beam, laser electron source a quadrupole triplet matches the electron beam and plasma cell are summarized in Table 1; those of the electron into the transfer line (see Fig. 3). Another quadrupole triplet just beam in Table 2. before the plasma is used to focus the beam into the plasma. Five additional quadrupoles are used to control the dispersion and the beta function. While the dispersion in the horizontal plane is almost zero along the part of the beam line downstream of the 2. The AWAKE beams: proton, electron, laser merging dipole (common line with the protons), the dispersion in the vertical plane is not closed due to the vertical kick given by the 2.1. Proton beam line tilted dipole, which merges the electron beam onto the proton beam axis. However, the final focusing system matches the beta For the main part of the beam line, the 750 m long CNGS functions and dispersion to the required 1σ spot size of r250 μm transfer line can be reused without major changes. However, in the at the focal point in both planes. Longitudinally the focal point is last 80 m a chicane has been integrated in order to create space for set at an iris (orifice) with a free aperture of 10 mm about 0.5 m a mirror of the laser beam line, necessary to merge the ionising upstream of the plasma cell. The present optics provides the laser pulse with the proton beam about 22 m upstream the plasma possibility to shift the focal point up to 0.8 m into the plasma cell cell. The proton beam is shifted horizontally by 20 mm at the without significant changes of the beam spot size. Ten kickers position of the laser mirror in the present layout [6]. Two beam (correctors) along the electron beam line compensate systematic position monitors (BPMs) and a beam loss monitor (BLM) are alignment and field errors and a shot-to-shot stability of 7100 μm placed around the merging mirror in order to interlock the SPS is predicted for a current fluctuation of 0.01% in the power con- extraction in case the mirror is hit by the proton beam. The syn- verters. In the common beam line upstream the plasma cell the chronization of the two co-propagating beams has to be stable to proton, electron and laser beam are travelling coaxially. Studies on 100 ps and the transverse pointing accuracy of the proton beam at the proton induced wakefields on the beam pipe walls and their its focal point is required to be r100 μm and r15 μrad, so that effect on the electrons show that the influence of the proton beam the proton trajectory is coaxial with the laser over the full length wakefields on the electrons is negligible [10]. However, direct of the plasma cell. Beam position monitors and screens (BTVs) in beam–beam effects show that the electron beam emittance blows

Please cite this article as: E. Gschwendtner, et al., Nuclear Instruments & Methods in Physics Research A (2016), http://dx.doi.org/ 10.1016/j.nima.2016.02.026i 4 E. Gschwendtner et al. / Nuclear Instruments and Methods in Physics Research A ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 3. Electron source and accelerating structure layout.

injection [12]. Combining the results of these studies show a Protons possible optimized injection scheme; the electron beam is offset in from SPS the common beam line and this offset is kept also at the injection point (see Fig. 5). Adding a kicker magnet close to the plasma cell Laser and focusing the beam into the iris allows to inject the electrons ﬁ Merging into the plasma wake eld with an offset of up to 3.25 mm and an Point angle between 0 and 8 mrad [9].

2.4. Laser beam line

The laser system is housed in a dust-free, temperature- Electron stabilized area (class 4) and includes the laser, pulse compressor Source and laser beam transport optics [13]. The laser beam line to the plasma cell starts at the output of the optical compressor in the laser lab and is transported via a newly drilled laser core. The laser beam line is enclosed in a vacuum system, which is attached to the compressor's vacuum chamber and to the proton beam line vacuum system at the merging point, which is at a vacuum level of 10 À7 mbar. The deﬂection of the laser beam will be performed with dielectric mirrors held by motorized mirror mounts, installed in the vacuum system. The size of the laser beam and the focusing spatial phase will be controlled by a dedicated telescope just before the pulse compressor. The focal distance is to be adjustable in the range between 35 and 45 m. Electron A diagnostic beam line will be installed in the proton tunnel for Merging Point measuring the beam properties of a low energy replica of the Common ionizing beam in exactly the distance which corresponds to the Beam Line plasma cell location. The laser beam for the electron gun will be taken from the second output of the base version of the laser system, further ampliﬁed to E30 mJ, compressed to 300 fs using an in-air compressor, frequency converted to 262 nm via a Third Harmonics Generation and stretched to the desired pulse length of 10 ps.

Start of 2.5. Low-level RF and synchronization Plasma Cell The proton, electron and the high power laser pulse have to arrive simultaneously in the rubidium plasma cell. The proton Fig. 4. The integration of the AWAKE beam lines (proton, electron and laser) in the bunch is extracted from the SPS about every 30 s and must be AWAKE facility (formerly CNGS). synchronized with the AWAKE laser and the electron beam pulsing at a repetition rate of 10 Hz. The latter is directly generated using a up [11]. In addition to the beam–beam effect, the most efﬁcient photocathode triggered by part of the laser light, but the exact electron injection into the plasma wake ﬁelds must be optimized. time of arrival in the plasma cell still depends on the phase of the The trapping of the electrons in the plasma was analysed with RF in the accelerating structure. Each beam requires RF signals at respect to the transverse vertical position y and angle y0 at characteristic frequencies: 6 GHz, 88.2 MHz and 10 Hz for the

Fig. 5. The 3σ envelope of the electron beam in the common beam line, in case the electron beam is coaxial with the proton beam (OnAxis) or with an offset to the proton beam axis (Offset). The position and free aperture of the main beam line elements are represented by blue squares. (For interpretation of the references to color in this ﬁgure caption, the reader is referred to the web version of this paper.)

rubidium vapor electron bunch r laser pulse modulated proton bunch plasma αi z proton bunch f defocusing trapped electrons laser pulse e region

Fig. 6. The oblique injection of the electrons into the plasma. synchronization of the laser pulse, 400.8 MHz and 8.7 kHz for the efficient electron trapping one needs the density ramp shorter SPS, as well as 3 GHz to drive the accelerating structure of the than E10 cm. To meet this requirement the vapor cell ends will electron beam [14]. A low-level RF system and distribution has now have a continuous flow through orifices at each end [13]. The been designed to generate all signals derived from a common Rb sources should be placed as close as possible to the orifices to reference. Additionally precision triggers, synchronous with the minimize the density ramp length. Thus there is continuous flow arrival of the beams, will be distributed to beam instrumentation of Rb from the sources to the plasma cell and afterwards from the equipment to measure the synchronization. Phase drifts of the plasma cell to the expansion volumes through the orifices (10 mm optical fibers transporting the RF signals for the synchronization of diameter). The walls of the expansion volumes should be cold the SPS with AWAKE will be actively compensated by newly enough (39 °C, the melting temperature of Rb) to condense all Rb developed hardware, which is essential to achieve a link stability atoms. The density gradient in the plasma cell will be controlled of the order of 1 ps. by the temperature difference of the Rb sources at each end of the plasma cell. In order to control a density gradient of 0.5–% with at least 50% precision the relative reservoir temperatures must be 3. Plasma source controlled with a precision of 0.1 °C or better.

AWAKE will use a rubidium vapor source [15] ionized by a short laser pulse (see Table 1). Rubidium plasma consists of heavy ions, 4. Electron injection that mitigate plasma ions motion effects. The plasma cell is 10 m long and has a diameter of 4 cm. The density uniformity is achieved In order to optimize the electron acceleration in the plasma by imposing a uniform temperature (within 0.2%) along the source. various schemes for electron injection have been investigated. The For that purpose a heat exchanger with sufficient heat carrying fluid history of the evolution on the optimized electron injection for flow is used. Synthetic oil is circulated inside a thermal insulation AWAKE is described in [16]. In the first experimental phase the around the tube containing the rubidium vapor. The oil temperature electron bunch will be at least one plasma period long in order to can be stabilized to 70.05 °C. A threshold ionization process for the avoid exact phasing with the proton bunch modulation and thus first Rb electron is used to turn the uniform neutral density into a cover several modulation cycles. Once the SMI is better under- uniform plasma density. The ionization potential is very low, stood and optimal parameters are found, it is planned to inject

ΦRb¼4.177 eV, as is the intensity threshold for over the barrier short electron bunches at the desired phase. 12 2 ionization (OBI), IIoniz 1:7 Â 10 W=cm . At the two ends of the 10 m long plasma cell fast valves were 4.1. Oblique injection foreseen originally, with the fast valves open only to let the proton, laser and electron beam pass. However, gas dynamic simulations At the plasma entrance the plasma density increases smoothly of the Rb vapor ﬂow showed that the fast valves were not fast from zero to the baseline density of 7 Â 1014electrons/cm3. Elec- enough (1–3 m) to ensure a short enough density ramp: for trons initially propagating along the proton beam axis are not

Please cite this article as: E. Gschwendtner, et al., Nuclear Instruments & Methods in Physics Research A (2016), http://dx.doi.org/ 10.1016/j.nima.2016.02.026i 6 E. Gschwendtner et al. / Nuclear Instruments and Methods in Physics Research A ∎ (∎∎∎∎) ∎∎∎–∎∎∎

160 5. Diagnostics 140 5.1. Direct SMI measurements 120

Ve The direct SMI diagnostic tools are based on transition radia-

G 100 /%, tion measurements [19]. oblique 80 side The optical transition radiation (OTR) is prompt and its time

egrah fl 60 structure re ects that of the bunch charge density hitting the radiator foil placed in the beam path. For AWAKE the plasma and C 40 – on-axis modulation frequency range is 100 300 GHz. The time structure of 20 the light and the proton bunch can be characterized using a ffi1ps resolution streak camera. The ability of the streak camera to detect 0 the ps modulation of the light signal was tested using beating laser 0 0.5 1.0 1.5 2.0 2.5 beams; the results show that the period of the modulation can be Energy,GeV measured for frequencies of up to 300 GHz [13]. Fig. 7. Final energy spectra of electrons in cases of side, on-axis, and oblique Coherent transition radiation (CTR) diagnostics uses time- injection methods. Beam loading is taken into account. resolved and heterodyne frequency measurements to determine the frequency and possibly amplitude of the SMI. The CTR has a trapped by the plasma wave in case the plasma density increases radially polarized electric field and has a maximum emission at a over a too long distance. The effect is similar to the plasma lens certain angle that depends on the particular experimental condi- effect [17] and is explained in detail in [16]. For the parameters of tions of the plasma density. This makes it difficult to couple into the AWAKE experiment, a transition region of 10 cm length is the fundamental mode of circular or rectangular wave-guides, sufficient to defocus the electrons. therefore using quasi-optical propagation to handle the CTR beam To shorten the transition area, the plasma cell ends are at least in the vicinity of the interaction point is considered. The designed with a continuous flow through orifices as described in CTR can be detected when focused onto a pyro-detector or to Section 3. With this design the defocusing region is on the order of couple part of a beam into a horn antenna and then detect it with the help of Schottky diodes. 15 cm. This distance is still sufficient to deliver a radial momentum of about 0.5 MeV/c to the electrons thus preventing their trapping 5.2. Indirect SMI measurement by the plasma wave. But fortunately the defocusing region does not extend beyond the radial plasma boundary [16]. So the elec- Protons in the defocusing phase of the wakefields exit the trons that are outside the ionized area in the plasma transition plasma close to the plasma center and appear as a narrow core on region propagate freely and some of them can even receive a small transverse profiles downstream the plasma. In [20] it is shown focusing push of several mrad. that the defocusing angle is on the order of 1 mrad for the Applying an oblique electron injection as shown in Fig. 6 400 GeV/c protons. For the indirect SMI measurements two beam- mitigates the loss of the electrons at the plasma density transition imaging screens (BTVs) at a distance of E8 m will be inserted region: the electrons have a small radial offset with respect to the downstream the plasma cell in order to measure the transverse proton beam upstream the plasma cell and are injected with a bunch shape and the beam size. Detailed studies of the screen material showed that a 1 mm thick Chromox-6 (Al O :CrO ) small angle αi into the plasma. In this way they approach the axis 2 3 2 in the region of already constant plasma density and therefore can scintillator with a hole or an OTR material at the beam centre can measure the beam shape and produce enough light, with only get trapped into the established plasma wave. The optimum values minimal interference of the proton beam. A defocused beam edge for the oblique injection scheme found in simulations are [16]: resolution of 0.6 mm can be achieved [20]. ξ ¼ : α ¼ : electron delay e 11 5 cm, injection angle i 2 8 mrad, and ¼ focusing point zf 140 cm. These values are compatible with the 5.3. Electron spectrometer range of settings for the electron beam line (see Section 2.3) and therefore no changes in the facility design are required. The electron spectrometer system [21] consists of a C-shaped dipole providing a 1.5 T field to separate the electrons from the proton beam and disperse them in energy onto a scintillating 4.2. Density gradient along the plasma screen (baseline is gadolinium oxysulfide). The spectrometer system also includes a quadrupole doublet in a point-to-point ima- With the plasma cell design as described in Section 3 a plasma ging configuration, to focus the beam exiting the plasma onto the density gradient of several percent along the 10 m long plasma cell spectrometer screen and increase the energy resolution. An optical can be created. If the gradient is positive the resulting change of line will transport the screen light to a CCD camera. Radiation the wakefield structure optimizes the electron acceleration. Details protection calculations showed that the CCD camera needs to be of these studies are described in [18]. Fig. 7 shows the final energy located at a position 17 m away from the screen in order to avoid of the electrons after passing through the 10 m long plasma cell for radiation damage. Calculations show that by transporting the light oblique, on-axis and side-injection. (In side-injection, the electrons to the camera using a series of mirrors and a suitable commercially available lens a peak signal to noise ratio of at least 120 is achieved would be injected into the plasma cell only after the SMI has at the expected electron acceleration capture efficiency. developed. This has turned out to be technically very challenging and was abandoned for the first phases of the experiment [6].) The results show that with oblique injection, realistic plasma bound- 6. The AWAKE underground facility and safety aspects aries at both ends and the linear growth of the plasma density by 1% over 10 m, about 40% of the injected electrons are trapped and Modifying the CNGS underground high radiation area for accelerated to E1.8 GeV after 10 m. AWAKE beam requires challenging modifications in a complex

Please cite this article as: E. Gschwendtner, et al., Nuclear Instruments & Methods in Physics Research A (2016), http://dx.doi.org/ 10.1016/j.nima.2016.02.026i E. Gschwendtner et al. / Nuclear Instruments and Methods in Physics Research A ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7 experimental area. AWAKE services and infrastructure must be order to separate the modulation of the proton bunch from the integrated within the existing radiation facility and at the same acceleration stage and to reach even higher gradients. time be designed and installed to keep the radiation dose to per- sonnel as low as possible. Shielding was installed and CNGS elements dismantled or exchanged in order to turn the high-radia- Acknowledgments tion, no-access CNGS facility into a supervised radiation area with safe regular access. Fig. 4 shows the integration of the AWAKE This work was supported in parts by: EU FP7 EuCARD-2, Grant beam lines in the CNGS facility; the diagnostics is downstream of Agreement 312453 (WP13, ANAC2); and EPSRC and STFC, United the plasma cell and not seen in the figure. Kingdom. The contribution of Novosibirsk team to this work is Fire safety equipment and the access procedures must be supported by The Russian Science Foundation, Grant no. 14-12- modified following extensive fire risk assessments, in order to deal 00043. The AWAKE collaboration acknowledges the support of with the specific AWAKE access needs (a stop-and-go proof-of- CERN, Max Planck Society, DESY, Hamburg and the Alexander von principle experiment) and properties of the new equipment. Humboldt Stiftung. Because of the high radiation, materials in CNGS were limited to concrete, iron, graphite, etc. leading to a very low fuel load of the area. The much lower radiation level in AWAKE means that racks, References electronics and other equipment with a non-negligible fuel load are still installed in an underground area that is 1 km distance [1] A. Caldwell et al., AWAKE Design Report, A Proton-Driven Plasma Wakefield from the nearest exit. In order to guarantee fire safety, a new and Acceleration Experiment at CERN, Internal Note CERN-SPSC-2013-013, CERN, more complex fire zone layout was created. Fire resistant walls Geneva, Switzerland, 2013. [2] R. Assmann, et al., AWAKE Collaboration, Plasma Physics and Control Fusion 56 and doors will ensure that at least two evacuation paths are (2014) 084013. available for every AWAKE area. Fire risk assessment were per- [3] N. Kumar, A. Pukhov, K. Lotov, Physical Review Letters 104 (2010) 255003. formed for the regions with the higher fuel loads (e.g. the oil- [4] A. Caldwell, K. Lotov, Physics of Plasma 18 (2011) 103101. fi [5] E. Gschwendtner et al., Performance and Operational Experience of the CNGS based Klystron for the electron source), re intervention and Facility, in: Proceedings of the 1st International Particle Accelerator Con- evacuation exercises are organised and the fire safety and eva- ference, IPAC2010, Kyoto, Japan, THPEC046, 2010. cuation information will be part of a dedicated AWAKE safety [6] C. Bracco et al., in: Proceedings of IPAC2014, Dresden, Germany, 2014. [7] K. Pepitone, S. Doebert et al., The electron accelerator for the AWAKE 436 course, obligatory for every person wanting to enter the area. experiment at CERN, EAAC2015, Elba, Italy, these proceedings, 2015. Radiation protection calculations have been performed using [8] G. Geschonke et al., CTF3 Design Report, Technical Report CTF3-Note-2002- FLUKA Monte Carlo code [22,23] to study the radiation environ- 047, CERN, Geneva, 2002. [9] J. Schmidt et al, Status of the Proton and Electron Transfer Lines for 441 the ment in the AWAKE facility. The results show that access to the AWAKE Experiment at CERN, EAAC2015, Elba, Italy, these proceedings, 2015. AWAKE experimental area has to be prohibited during the proton [10] U. Dorda et al., Simulations of Electron–Proton Beam Interaction before Plasma beam operation as the prompt dose equivalent rate exceeds in the AWAKE Experiment, IPAC15, Richmond, USA, 2015. 100 mSv/h. As a consequence, the experiment and equipment [11] U. Dorda, et al., Propagation of electron and proton beams before the plasma (2015) https://indico.cern.ch/event/403300/. must be remotely controlled. To allow partial access to the laser [12] A. Petrenko et al., Latest electron trapping simulation results, 2015, https:// room and the klystron area during the operation of the electron indico.cern.ch/event/403300/. beam, an appropriate shielding wall around the electron gun has [13] A. Caldwell at al., AWAKE Status Report, CERN-SPSC-2015-032, SPSC-SR-169, 2015. to be designed. During electron operation radiation levels in [14] M. Bernardini, et al., AWAKE, a Proof-of-Principle R&D Experiment at CERN, accessible areas will be continuously monitored. Moreover, a IPAC15, Richmond, USA (2015). dedicated air management system of the area including an air- [15] E. Öz, F. Batsch, P. Muggli, An acccurate Rb density measurement 454 method for a plasma wakefield accelerator experiment using a novel Rb 455 reservoir, borne radioactivity monitor has to be installed to guarantee safe EAAC2015, Elba, Italy, these proceedings, 2015. access conditions after proton beam operation. [16] A. Caldwell and AWAKE Collaboration, Path to AWAKE: Evolution of 457 the concept, EAAC2015, Elba, Italy, these proceedings, 2015. [17] P. Chen, Particle Accelerators 20 (1987) 171. [18] A. Petrenko, K. Lotov, A. Sosedkin, Numerical Studies of Electron Ac461 cel- 7. Summary eration Behind Self-Modulating Proton Beam in Plasma with a Density 462 Gradient, EAAC2015, Elba, Italy, these proceedings, 2015. [19] P. Muggli, AWAKE, proton-driven plasma wakefield experiment at 464 CERN, AWAKE is a proof-of-principle accelerator R&D experiment IPAC15, Richmond, USA, 2015. currently being built at CERN. It is the first proton-driven wake- [20] M. Turner et al., Indirect Self-Modulation Instability Measurement Concept for field acceleration experiment worldwide, with the aim to provide the AWAKE Proton Beam, EAAC2015, Elba, Italy, these proceedings, 2015. a design for a particle physics frontier accelerator at the TeV scale. [21] L.C. Deacon et al., Development of a spectrometer for proton driven plasma wakefield accelerated electrons at AWAKE, IPAC15, Richmond, USA, 2015. The installation of the AWAKE experiment is advancing well. [22] G. Battistoni, S. Muraro, P.R. Sala, F. Cerutti, A. Ferrari, S. Roesler, A. Fasso, J. Hardware and beam commissioning is planned for the first half of Ranft, in: M. Albrow, R. Raja eds., Proceedings of the Hadronic Shower – 2016. The physics of the self-modulation instability as a function of Simulation Workshop 2006, Fermilab 6 8 September 2006, AIP Conference Proceedings, vol.896, 2007, pp. 31–49. the plasma and proton beam properties will be studied starting by [23] A. Ferrari, P.R. Sala, A. Fasso, J. Ranft, FLUKA: a multi-particle transport code, the end of 2016. The longitudinal accelerating wakefield will be CERN-2005-10, INFN/TC-05/11, SLAC-R-773, 2005. probed with externally injected electrons starting by the end of 2017. At a later stage it is foreseen to have two plasma cells in

Please cite this article as: E. Gschwendtner, et al., Nuclear Instruments & Methods in Physics Research A (2016), http://dx.doi.org/ 10.1016/j.nima.2016.02.026i